Computational chemistry and mechanistic modeling of atmospheric chemistry: Models of alkane and oxygenate chemistry

Atmospheric pollutants such as ozone from volatile organic compound (VOC) and nitrogen oxide (NOx) interactions are deleterious to human health. Since VOCs are a major contributor to increased ozone levels observed in the atmosphere, VOC emissions have been limited by the Environmental Protection Agency, and industries have strived to avoid emissions of VOCs and use VOCs that are less detrimental to the atmosphere. The focus of this research was to develop an automated mechanism generation framework to create mechanisms for atmospheric chemistry of VOCs and NOx, that can predict the ozone formation potential of single components and mixtures of VOCs.; The reactions of VOCs in the troposphere were classified according to a small set of reaction families based on the literature, and mathematical operators were formulated to implement these reaction types in automated mechanism generation. A hierarchy was developed that enables the rate constant for every reaction to be specified as it is generated. An object-oriented programming framework was built to handle the diverse types of rate constants that characterize atmospheric reaction rates and to incorporate experimental data for reaction rate parameters seamlessly when it was available. Structure/reactivity relationships relying on estimates of thermodynamic properties were developed and used to estimate unknown reaction rate constants. Many of the radicals present in atmospheric reaction mechanisms have no experimental thermodynamic or group additivity data available. To fill the voids in this critical data high-level quantum chemical calculations were carried out to estimate thermochemical group additivity parameters for species relevant to the atmosphere. A method for estimating photolysis rate constants by estimating the absorption cross sections using a group additivity scheme was developed. This is the first demonstration of using group additivity to calculate absorption cross sections. Finally, automated mechanism generation and its associated algorithms for rate constant specification were applied to single components and VOC mixtures of alkanes, aldehydes, and ketones. The model results compared very well to experimental data, and the models outperformed other models in the literature.